It appears that what is happening is PL/pgSQL is caching the table
definition (it appears to do this on first execution), testing it with
dynamic SQL via the EXECUTE clause doesn't exhibit the same issue:
CREATE OR REPLACE FUNCTION test_fxn() RETURNS SETOF RECORD AS $$
DECLARE
test_rec RECORD;
On Tuesday 16 January 2007 10:10, "Lenorovitz, Joel"
<[EMAIL PROTECTED]> wrote:
> Greetings,
>
> I am trying to work with a TEMP TABLE within a plpgsql function and I
> was wondering if anyone can explain why the function below, which is
> fine syntactically, will work as expected the first time i
On Tue, Jan 16, 2007 at 11:10:25AM -0700, Lenorovitz, Joel wrote:
> Greetings,
>
> I am trying to work with a TEMP TABLE within a plpgsql function and I
> was wondering if anyone can explain why the function below, which is
> fine syntactically, will work as expected the first time it is called,
>
Greetings,
I am trying to work with a TEMP TABLE within a plpgsql function and I
was wondering if anyone can explain why the function below, which is
fine syntactically, will work as expected the first time it is called,
but will err out as shown on subsequent calls. The DROP TABLE line
seems to
